The camera never lies

December 17, 2008

If a picture is worth a thousand words, what happens when your computer can give those words a rewrite, mix them with other passages and generally treat them like newspaper scraps on a ransom note?

Eventually we won’t need a file system to browse our pictures – we will probably virtually walk through the places we’ve been with our photos like paintings on the wall of a gallery (with those of others there too). We are also getting closer to the day where images can be manipulated at a higehr level than the manual, pixel by pixel way we are used to: “I’d like a field with four golfers in it, overlooking the sea. Great, just pan left 30% and make the grass greener. Done.” With a large enough archive of images and the kind of technologies below, that scenario is not far off.

Here are a couple of great examples which can already be used and one more on the horizon.

Seadragon/Photosynth

“You familiar with Photosynth?”

“Yes… Taking a large collection of photos, analysing the similarities and displaying them in a reconstructed three dimensional space?”

“Exactly. Build me a high school gym.”

Developer conference? Nope. This is from Microsoft Photosynth’s airing in an episode of CSI earlier this year. As Microsoft put it, Photosynth is a perfect example of a tool creating a 1 + 1 = 5 scenario, where the thousands of pictures uploaded to the likes of Flickr can be combined to create a seamless three dimensional environment. Try it.

The cool news? The technology on which Photosynth is based just hit the iPhone. It’s called Seadragon, was built by a small Seattle area startup acquired by Microsoft in 2006, and allows you to seamlessly zoom in and out of a gigapixel scale image. Try it out in your browser or check out the iPhone demo.

Also check out Blaise Aguera y Arcas’s groundbreaking demo of Photosynth at TED 2007

Source

…and on CSI!

Seam carving: resizing no longer considered harmful.

contentawareresizing

This is an awesome technology originally developed by Israeli researchers Shai Avidan and Ariel Shamir. When you resize an image normally, you squash everything in it. With seam carving you only touch the parts that matter least. In other words, the golfers stay fat, but the sky and grass around them gets progressively removed.

This video by the researchers is the best way of getting a handle on the possibilities: the principle of seam carving also makes it possible to enlarge images without stretching them and selectively remove parts of a picture (e.g. you could pick a golfer and make them disappear more convincingly than traditional techniques). Check it out below. The paper also makes good reading.

The cool news: it’s in Photoshop CS4, which was released in October (yes, I know GIMP had it first). They call it “Content Aware Scaling”, and have a video demo here – (yep, the golfers).

Talking of Adobe: Infinite Images

On a continuing theme of algorithmically creating new images from the old, Adobe are playing around with a tool which takes any picture, finds other similar pictures and seamlessly stitches them into an infinitely pannable and zommable virtual environment.

Essentially, this is Photosynth, but instead of sticking to images of the same thing it grabs anything that fits the bill. Grainy footage from Adobe MAX 2008, via ReadWriteWeb.

What will happen when anyone can mash up two images and create a picture of a place that looks absolutely real? What other technologies are getting us closer to this world?

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: